Who this is wrong for.

We do not write courses. There is no syllabus of ours sitting behind this, no library of stock lessons to fill the gaps with, and nobody here who knows your subject better than you do. The conversion inherits exactly what you send: if a chapter is four pages of headings and a diagram, what comes back is four pages of headings and a labelled gap where the diagram was.

Providers get this wrong in good faith, because the binder feels substantial when it is sitting on a shelf. The honest test is to open the chapter you teach most often and ask whether a student who had only that chapter could pass. If the answer is no, what you need is a writer, and we are not one.

Those exist, there are dozens of them, they cost about fifty euros a month, and several are genuinely good. If what you want is a place to upload lessons you have already written, buy one of those and do not pay us. The work we do is the conversion — reading your book, matching every claim to the page it came from, and lifting your questions out without changing them — and if you are prepared to do that yourself, the software around it is not worth a few thousand euros.

The trap is the one nobody names on the pricing page. You pay fifty euros, you get an empty account, and the work is still sitting on your desk in the same binder it was in last year. We have no figure for how often that is how it ends, and anybody who quotes you one has invented it. Ask yourself instead which evening this month you were going to spend typing your book in. That is not a criticism of the software.

Your questions go in as you wrote them, or they do not go in. We will not modernise the wording, split a two-part question, fix a comma, or replace a scenario about a fax machine with one about a smartphone. If you ask us to, we will say no, and if that is a problem then we are not the right supplier — this is the one refusal on the page we would not bend for a large order.

The reason is not stubbornness. Fifteen years of past papers are worth something precisely because they are the questions your students were actually asked and your teaching was actually built around. Improving them produces a bank that is nobody’s in particular, which is the thing you can already buy for nothing.

We ask who owns the copyright before we convert a page, and it is the question that ends the most conversations. A great deal of licence-training material is a publisher’s standard, an awarding body’s workbook, or a manual bought in bulk from a supplier with a licence that covers printing it and nothing else. Teaching from it is not the same permission as digitising it.

If you hold written permission, send it and we will read it. If you do not, we decline the work. That is not us being careful on your behalf: converting somebody else’s book into a product you sell puts you in front of the rights holder, not us, and we are not going to take your money to arrange that.

A chapter comes back within five working days. A whole book is a project measured in weeks, and most of that time is a person reading the output rather than a machine producing it. There is no configuration of this that produces a finished course in seven days, and a supplier who tells you otherwise is describing a machine translation of your PDF with a quiz stapled to the end.

If a deadline is already on top of you, the honest advice is to teach the term the way you taught the last one and start this in the quiet month. We would rather have that conversation now than deliver something late and be the reason.

That is a different product, it is widely available, and much of it is free. Type a description of your subject and a language model writes lessons and questions about it. What comes out is fluent, confident, and unattached to any document you own. It is also, for a regulated licence, unusable — nobody can tell you which page a claim came from, because there was no page.

No language model is called anywhere in this product. Not as a boast, and not permanently: the four origin labels exist precisely so that if any part of a lesson is ever machine-written, the paragraph says so where your students can see it. What we will not do, then or now, is produce a question we cannot point at a page for.

We will not quote one. There is no pilot behind us large enough to support a number like that, and in a business where you answer to a regulator, an unsupported claim about results is worse than no claim at all — it is the sentence that gets read back to you. Any supplier quoting you an improvement figure with no named school behind it has made it up, and you should ask which school.

We are also not going to promise that your students will use it. We have no usage data, because we have no users. What we can tell you is what was removed from the path: no app store, no download, nothing to install.